**4. Discussion**

In this study, we used a multi-method approach to (i) test a hypothesis that each method separately may be useful in donor characterization and to (ii) try to distinguish a regular donor of feces from the stool bank (as we know his stool is beneficial for patients and he is generally healthy and meets all inclusion criteria for being a stool donor) from randomly chosen persons.

Although very complex, our analysis should be still considered as a preliminary one with several limitations. First of all, we tested only three donors in total in our study; however, our goal was to perform complex analysis with screening of multiple samples per patient in a time-resolved manner. Additionally, a relatively weak point of our analysis could also be using a limited number of media for the cultivation experiment; however, we carefully chose them to cover both aerobic and anaerobic taxa.

Taking into account all limitations, our work showed that assessment of the intestinal microbiota is very complex. Each of the applied methods separately had significant limitations, however, enabled some significant observations. Classic bacterial culturing showed that single intestinal microbiota sampling can give false, and surely not representative, results, because many bacteria are found in the feces irregularly, and the composition of fecal microbiota may, to some extent, change daily [35]. For a complete picture of the composition of the intestinal microbiota, multiple culturing should be performed, and obtained results should be analyzed together. The next applied procedure, i.e., flow cytometry, showed that live bacteria dominated the samples. However, discussion on the active component of the intestinal microbiota preparation continues. It is said that live bacteria are not necessarily responsible for the beneficial properties of intestinal microbiota, and metabolites or even individual genes should be taken into consideration [36]. Nevertheless, the prevalence of living bacteria is a positive aspect in the argumen<sup>t</sup> that live bacteria are the factor that restores intestinal homeostasis [37]. The last of the applied methods, i.e., next-generation sequencing, definitely brings the most abundant data, considering the bacterial diversity of the intestinal microbiota. The weakness of this method, however, is that the results are relative; we do not have absolute numbers of individual genera and the bacteria represented in a very low titer may be underrepresented in the sample [38]. Keeping in mind that this is a semi-quantitative method, not a quantitative one, and inferences based on the indicated percentages must be careful, we can see the differences that we observed between donor C and donors A and B may point to unique combinations of individual species. Possibly, the key species for health processes are present in donor C, who is a regular donor of feces. Such conclusions will always be arbitrary, but the "beneficial" species described in the literature, such as *Blautia*, *Barnesiella*, *Butiricimonas* [18] and *Roseburia*, are significantly more abundant in donor C than in others; however, other strains considered beneficial, such as *Bacteroides* and *Faecalibacterium*, occur in comparable percentages (Figure 5).

More definite conclusions can be drawn when looking collectively at the results obtained with the application of all these methods. In general, the metabarcoding analysis revealed that Shannon's

biodiversity index was significantly higher for donor C than others. This is in agreemen<sup>t</sup> with previous results, indicating that biodiversity is one of the most important markers of a healthy microbiome [39,40]. There are no specific values of diversity indexes that particular samples of intestinal microbiota should achieve in order to be recognized as more valuable than others. However, high biodiversity indicates a richer intestinal flora and is associated, according to current knowledge, with health [14].

Regarding a microbiota composition assessed by classical culturing and NGS, we found some species and genera being unique for donor C, and it was also found that abundances of particular species may play a role. The donor C microbiota was most stable in time, which may also be an indicator of a super-donor. However, even in donor C's stool samples, we observed some day-to-day diversity. A possible lack of influence of this variability on overall microbiotal homeostasis may be explained by the hypothesis that a healthy "functional core" of the microbiome is maintained by a complement of metabolic and other molecular functions that are performed by the microbiome within a particular habitat but are not necessarily provided by the same organisms in di fferent individuals [41]. Therefore, according this hypothesis, the microbiome as a pool of genes remains stable.

In this study, we identified genera (*Roseburia, Barnesiella, Blautia, Butiricimonas*) considered beneficial as more abundant in donor C than in the others. ANCOM analysis on samples from donor A versus C highlighted *Acidaminococcus* and *Paraprevotella* to be more abundant in samples from donor C. Both of these genera are common human gu<sup>t</sup> microbiota, with the former shown to be able to use glutamic acid as the only source of energy and produce hydrogen and hydrogen sulfide from citrate [42,43]. *Paraprevotella*, on the other hand, is an opportunistic human pathogen, associated with non-alcoholic fatty liver disease [44,45]. The same analysis has shown *Anaeroplasmatales* and *Gastranaerophilales* orders to be more abundant in samples from donor A than from donor C. While not much is known about the *Gastranaerophilales* order, *Anaeroplasmatales* is a Gram-negative obligatory anaerobe, a member of Mollicutes class, which also harbors some human pathogens such as *Mycoplasma* and *Ureaplasma*.

The same analysis of the second pair (donor B versus C) also highlighted *Anaeroplasma* in this setting to be more abundant in donor B, along with the *Holdermanella* genera, a Gram-positive member of the Firmicutes phylum associated with chronic kidney disease [46]. As for taxa more abundant in the donor C microbiota, two members of the Firmucutes phylum were detected: *Lachnospiraceae* and *Dialister*. Members of the *Lachnospiraceae* family are common human gu<sup>t</sup> microbiota and have been linked to protection from colon cancer (and obesity from the other side), due to their ability to produce butyric acid, which is an important factor for both microbial and host epithelial cell growth [47]. *Dialister* is a Gram-negative member of the Negativicutes class, usually associated with human dental infections [48,49]. We found that *Anaeroplasma* is more abundant in donors A and B, and positively correlates with SYTO9-, PI- cell counts (double negative, which may be considered as not-alive-not-dead). Could those be markers of "good" and "bad" donors? To answer this question, numerous population studies should be performed correlating, e.g., *C. di*ffi*cile* infection remission with particular donors and particular core microbiota as well as unique bacteria present in their stool.

By comparing flow cytometry with amplicon analysis results, we found a positive correlation between the double negative cell group and *Anaeroplasma* genera. It is possible that, due to their unusual, rich in sterols, cell membrane, members of Mollicutes class, such as *Anaeroplasma*, can be attributed to high numbers of SYTO9-, PI- cells. However, they are definitely not the only ones that built this group, because the relative abundance of Mollicutes was very small in samples from donor C, while SYTO9-, PI- cells per sample percentage in this donor was comparable to others. Nonetheless, there is a possible pattern that high relative abundance of *Anaeroplasma* is more characteristic of donors A and B, being "non-donors" than of donor C, a regular stool donor. It is also possible that double negative cells are bacterial spores.

The selection of an appropriate stool donor is a key step in FMT [50]. Donors are thoroughly screened to ensure they do not harbor any transmittable pathogens or disease. Comparing the gu<sup>t</sup> microbiota profiles of di fferent donors, it is known that microbial diversity is a reliable predictor of FMT success; however, a variety of additional factors, both genetic and environmental, are also known to influence FMT results [51]. As we showed in this work, a regular clinical questionnaire should be coupled with multiple tests results, as only together can they give reliable results, and they will definitely give more data than each method separately.
